AI Daily Insights No. 95 · November 7, 2026

AI Tracker

Follow AI companies and products through sourced briefs and linked editorial coverage.

Source publication time · Beijing time

Checked every 4 hours · Last checked (Beijing) · 0 new in this 4-hour window

All updates

29 updates in the archive

TechCrunch reports Box will build Nvidia Nemotron into its content retrieval system

According to TechCrunch, Box showed off Box Content Retrieval at its developer conference, a new feature that embeds Nvidia’s Nemotron system into the core of its product. The company says the agent-first retrieval layer should help queries return faster, more accurate, and less expensive results. Nvidia VP of generative AI software Kari Briski said queries are expected to be answered four times faster, with 20% fewer errors and 90% fewer tokens used. The feature will roll out early next year to Box Enterprise Advance customers.

AI-generated Announced TechCrunch ↗ #Nvidia

GitHub Copilot enters Xcode’s built-in coding assistant in public preview

GitHub said on the GitHub Blog that GitHub Copilot is now available directly inside Xcode’s built-in coding assistant, starting as a public preview in Xcode 27.3 beta. The integration gives developers another way to use Copilot in a native Xcode workflow, while the existing GitHub Copilot for Xcode extension remains available as a separate option. GitHub says it supports all Copilot plans, and setup requires Xcode 27.3 beta on a Mac running macOS Tahoe 26.6 or later.

AI-generated Preview GitHub Copilot ↗ #Apple#Microsoft

The Verge reports OpenAI rolls out Images 2.6 in ChatGPT, Codex and the API

According to The Verge, OpenAI is rolling out Images 2.6 across ChatGPT, Codex and its API, saying image generation is now 50% faster. The update also adds native GIF creation from a prompt or an uploaded photo in ChatGPT, with controls for playback speed and frame-by-frame editing. OpenAI says the rollout is reaching ChatGPT and Codex users on web, desktop and mobile, while the API gets GPT-Image-2.6 Flare and Sunburst for faster, higher-quality image generation.

AI-generated Released The Verge ↗Techmeme ↗ #OpenAI#Codex#ChatGPT
Read full coverage

Bloomberg says Waymo plans to end its exclusive Uber robotaxi deal in 2028

According to Bloomberg, Waymo plans to end its exclusive robotaxi partnership with Uber in Austin and Atlanta in January 2028 and move those rides to its own app. Waymo co-CEO Tekedra Mawakana said the change responds to customer feedback asking for a more consistent ordering experience. For now, riders in those two U.S. cities can hail Waymo vehicles only through Uber’s platform, and the exclusive arrangement remains in place until then.

AI-generated Planned Bloomberg ↗Techmeme ↗ #Alphabet#Waymo#Uber

GitHub separates bring-your-own-key usage in Copilot reports

GitHub updated its Copilot usage API so organizations can distinguish usage of models supplied by developers or administrators from models hosted by GitHub. Requests configured locally in the Copilot CLI or app go directly to the developer’s provider or local model. Requests for enterprise custom models pass through GitHub. Reports list a custom model separately even when its name matches a hosted model. The breakdown appears in one-day and 28-day reports at both user and aggregate levels.

AI-generated Released GitHub ↗ #Microsoft

NVIDIA describes an AI-defined fab for faster semiconductor engineering

NVIDIA described AI applications in semiconductor fabs spanning engineering simulations, process prediction and testing factory changes in virtual environments. According to the company, COMSOL uses cuDSS to accelerate multiphysics simulations by up to 10x, while Applied Materials uses PhysicsNeMo to accelerate topography simulations by up to 35x. Siemens is integrating cuLitho into its Calibre lithography software, delivering up to 17x acceleration on key simulation workloads. The faster calculations allow engineers to verify more before a wafer goes to print.

AI-generated Technical overview NVIDIA ↗ #Nvidia

JetBrains Air adds GitHub Copilot for local and cloud coding tasks

Developers can now select GitHub Copilot in JetBrains Air to delegate coding work, follow parallel sessions and review changes alongside the code. The integration uses the Agent Client Protocol and supports an existing Copilot subscription. Local tasks run through the IDE. For cloud work, the business organization needs Air Teams and an administrator must enable Copilot. Each cloud task gets a container on a fresh virtual machine and consumes Copilot plan usage rather than JetBrains AI credits.

AI-generated Released GitHub ↗ #Microsoft

Cisco uses OpenAI models to connect signals across its network

Cisco described how its security teams combine network telemetry with OpenAI models to identify activity that warrants investigation. Two signals that look harmless on their own can reveal a pattern when examined together across the network. Cisco also distributes access controls across microservices, containers and clusters, restricting communication to the connections an application needs. During an investigation, defenders can isolate the affected systems more selectively and allow other services to keep operating.

AI-generated In use OpenAI ↗ #OpenAI

OpenAI describes Heimstaden using ChatGPT Enterprise and Codex for property management

OpenAI published a case study on European property operator Heimstaden using ChatGPT Enterprise and Codex for letting prices, property listings and environmental-risk analysis. According to the case study, its Dutch team reduced the cost of a virtually staged listing image from about €50 to less than €0.10, with estimated annual savings potential of €130,000 in the Netherlands. Analysts also used Codex to process environmental-risk data for 20,112 assets, producing more than six million useful data rows. Employees retain review and approval of business decisions.

AI-generated Technical overview OpenAI ↗ #OpenAI#Codex#ChatGPT

NVIDIA describes physical AI and agentic AI deployments for city services worldwide

NVIDIA described physical AI and agentic AI applications in city services in Spain and Malaysia. According to the company, 1MillionBot combines ALIA, Nemotron and other models for more than 150 deployments in Spain, handling municipal procedures, appointment requests and service inquiries. In Malaysia, ITMAX plans to scale its SamurAI V2 vision agents to support more than 17,000 cameras by 2027 and uses traffic simulation to test signal timing. NVIDIA says the approach can reduce peak-hour corridor travel times by up to 26%.

AI-generated Technical overview NVIDIA ↗ #Nvidia